mlsanityData to 7 Oct 2026

Mini-BEHAVIOR-Gran: Revealing U-Shaped Effects of Instruction Granularity on Language-Guided Embodied Agents

arXiv 2604.17019 ↗First posted 18 Apr 2026cs.AIcited by 1

Not on the map. Why?

Abstract (arXiv)

Instruction granularity is an important yet poorly controlled variable in language-guided embodied AI. Existing benchmarks typically pair each task with a single static instruction, making it difficult to study how agent behavior changes when the same task is described at different levels of detail. We introduce Mini-BEHAVIOR-Gran, a new benchmark for controlled studies of instruction granularity that extends Mini-BEHAVIOR with multiple instruction variants per task, ranging from high-level goal descriptions to step-by-step guidance. Using this benchmark, we compare four candidate metrics for cross-task granularity quantification: token count, entity count, action-verb count, and planning-width, and find that width correlates most consistently with agent performance. Using width to organize training and evaluation further reveals a non-monotonic U-shaped relationship between instruction granularity and performance, with peaks at both fine and coarse extremes. Further analysis suggests that the coarse-granularity performance rebound is associated with shallow grounding, where agents learn vision-dominant policies.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 38 of its 56 references that are in the map. First posted on arXiv 18 Apr 2026.

Builds on

38 of 56 references.

References · 38

  1. 2023Sigmoid Loss for Language Image Pre-Trainingcited by 3,048
  2. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 2,993
  3. 2023ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIscited by 1,880
  4. 2025π0.5: a Vision-Language-Action Model with Open-World Generalizationcited by 1,653
  5. 2023Gorilla: Large Language Model Connected with Massive APIscited by 1,302
  6. 2021ViLT: Vision-and-Language Transformer Without Convolution or Region Supervisioncited by 1,224
  7. 2025GR00T N1: An Open Foundation Model for Generalist Humanoid Robotscited by 1,167
  8. 2022Inner Monologue: Embodied Reasoning through Planning with Language Modelscited by 1,057
  9. 2025Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Successcited by 844
  10. 2019ALFRED: A Benchmark for Interpreting Grounded Instructions for Everyday Taskscited by 773
  11. 2023An Embodied Generalist Agent in 3D Worldcited by 355
  12. 2024Prismatic VLMs: Investigating the Design Space of Visually-Conditioned Language Modelscited by 336
26 more
  1. 2023Minigrid & Miniworld: Modular & Customizable Reinforcement Learning Environments for Goal-Oriented Taskscited by 320
  2. 2024A Survey on Vision-Language-Action Models for Embodied AIcited by 274
  3. 2025Motus: A Unified Latent Action World Modelcited by 231
  4. 2021Benchmarking the Spectrum of Agent Capabilitiescited by 193
  5. 2021iGibson 2.0: Object-Centric Simulation for Robot Learning of Everyday Household Taskscited by 191
  6. 2025ThinkAct: Vision-Language-Action Reasoning via Reinforced Visual Latent Planningcited by 152
  7. 2020Watch-And-Help: A Challenge for Social Perception and Human-AI Collaborationcited by 131
  8. 2018Coarse-to-Fine Decoding for Neural Semantic Parsingcited by 110
  9. 2025A Survey on Vision-Language-Action Models: An Action Tokenization Perspectivecited by 96
  10. 2025Discrete Diffusion VLA: Bringing Discrete Diffusion to Action Decoding in Vision-Language-Action Policiescited by 90
  11. 2025PD-VLA: Accelerating Vision-Language-Action Model Integrated with Action Chunking via Parallel Decodingcited by 88
  12. 2025InstructVLA: Vision-Language-Action Instruction Tuning from Understanding to Manipulationcited by 71
  13. 2025BitVLA: 1-bit Vision-Language-Action Models for Robotics Manipulationcited by 42
  14. 2025MPO: Boosting LLM Agents with Meta Plan Optimizationcited by 36
  15. 2018Learning Features and Abstract Actions for Computing Generalized Planscited by 35
  16. 2025Shortcut Learning in Generalist Robot Policies: The Role of Dataset Diversity and Fragmentationcited by 31
  17. 2025ReflAct: World-Grounded Decision Making in LLM Agents via Goal-State Reflectioncited by 30
  18. 2023Mini-BEHAVIOR: A Procedurally Generated Benchmark for Long-horizon Decision-Making in Embodied AIcited by 23
  19. 2025AVA-VLA: Improving Vision-Language-Action models with Active Visual Attentioncited by 22
  20. 2017Accurately and Efficiently Interpreting Human-Robot Instructions of Varying Granularitiescited by 19
  21. 2023ThinkBot: Embodied Instruction Following with Thought Chain Reasoningcited by 19
  22. 2021Draw Me a Flower: Processing and Grounding Abstraction in Natural Languagecited by 14
  23. 2025MuSC: Improving Complex Instruction Following with Multi-granularity Self-Contrastive Trainingcited by 12
  24. 2026Restoring Linguistic Grounding in VLA Models via Train-Free Attention Recalibrationcited by 11
  25. 2023General Policies, Subgoal Structure, and Planning Widthcited by 6
  26. 2020Domain Adaptation for Semantic Parsingcited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026JRDB-AVR: An Active Visual Reasoning Benchmark for Embodied Agents in Real-World Environmentscited by 0